关于“Using custom containers with AI Platform Training”的评价

Anna Elisabetta Z. · 评论over 3 years之前

Jerome V. · 评论over 3 years之前

Rafael G. · 评论over 3 years之前

Manikant J. · 评论over 3 years之前

Tony R. · 评论over 3 years之前

NameError: name 'df_train' is not defined !bq extract \ --destination_format CSV \ covertype_dataset.training \ $TRAINING_FILE_PATH Waiting on bqjob_r5a514c61aa269f4c_0000018091715755_1 ... (0s) Current status: DONE

Tony R. · 评论over 3 years之前

NameError: name 'df_train' is not defined when splitting my train/val data sets it appeared as though the train file was missing, quicklabs accepted the checkpoint for splitting dataset. When I went to convert numerics to float64 "NameError: name 'df_train' is not defined" became the issue

Tony R. · 评论over 3 years之前

Fayyaz F. · 评论over 3 years之前

ALAN S. · 评论over 3 years之前

Kiran R. · 评论over 3 years之前

sachin s. · 评论over 3 years之前

2 hours was not enough, processes took very long to start or to complete. Checkpoints did not turn to complete even though the task was clearly done and process was executed on the cloud. It took 10-15 minutes until I was able to mark it as done.

Georgios G. · 评论over 3 years之前

Renuka H. · 评论over 3 years之前

Tondam N. · 评论over 3 years之前

sachin s. · 评论over 3 years之前

José Luis G. · 评论over 3 years之前

TieSheng W. · 评论over 3 years之前

Shinichi M. · 评论over 3 years之前

Alvin C. · 评论over 3 years之前

Very good lab to understand e2e ml project. One drawback is consistency of the lab instruction; e.g. training split takes 80% of sample, but it actually takes 40% according to the query.

Nattacha P. · 评论over 3 years之前

This was a good one, but the time was tight. There wasn't much slop with just copy pasting code from the reference notebook. I expect many/most people would run out of time.

Guy M. · 评论over 3 years之前

Time is not sufficient

Ruchi A. · 评论over 3 years之前

Time is not sufficient

Ruchi A. · 评论over 3 years之前

John N. · 评论over 3 years之前

John N. · 评论over 3 years之前